W3C home > Mailing lists > Public > public-rif-wg@w3.org > July 2008

Re: [PRD] is in a freezable state

From: Christian de Sainte Marie <csma@ilog.fr>
Date: Fri, 04 Jul 2008 10:20:03 +0200
Message-ID: <486DDD33.50007@ilog.fr>
To: Gary Hallmark <gary.hallmark@Oracle.com>
CC: RIF WG <public-rif-wg@w3.org>


Gary Hallmark wrote:
> 2.1.1, (maybe elsewhere, too) contain the following sentence -- 
> It is specified in the normative schema as a substitution group.
> I would delete this sentence because
> a. there is no normative schema
> b. it probably isn't a substitution group.  BLD uses a group element, 
> not a substitution group.

Right. I forgot to remove that. I will do it.

> Should also log an ISSUE about how the BLD and PRD schemas should 
> interrelate -- should they both include a Core schema so we are sure 
> about what the syntactic subset is, or are the schemas separate and we 
> have to check by hand that instances overlap as intended?

I will do it.

> 2.2 -- RIF-PRD defines one single abstract class for actions: ACTION, 
> that is realised by three concrete constructs:
>    * the ASSERT and Retract constructs are associated with an Atom or a
>      Frame that represents the target of the actions.
> It isn't clear that the 3 concrete constructs are Retract, Atom, and Frame

Ooops! Remnant of past times, when we used to have explicit Assert, 
Retract and Execute. I will correct that sentence.

>, editor's note -- we already have a way to make an assertion PRD 
> or BLD specific.  E.g.
> P(?x) :- Q(?x)  // core
> Do(P(?x)) :- Q(?x)  // PRD only
> And(P(?x)) :- Q(?x) // BLD only

The question is meant to be the other way around: do yo uwant an 
explicit Assert? I will rephrase the note.

> 2.4 and elsewhere in 2.*
> I think you should change <Var>?c</Var> to <Var>c</Var>
> unless you mean that it should be ??c in the PS

If the name of the var is ?c in the original rule, it should be 
<Var>?c</Var> in the XML. I already commented several times that the PS 
should not change the name of the variables, and I had exactly this case 
in mind.

So, yes, I mean that the name of the variable should be ?c including in 
the XML. If that means ??c in the PS, let's do that, or change the PS.

I will post, separately, a suggestion to that effect (that is, change 
the PS).

> 3.4.
> What are the semantics of ground actions in a ruleset?  I think these 
> could be modeled as an initial instantiation.

In the current draft, they are processed like the other rules. If enough 
of the PR languages that PRD intends to cover have initialisation 
constructs, we will consider adding one in a future version of the draft.

My personal view is that PR are not meant nor used to represent 
facts,and, thus, there is no reason to have a specific feature for that 
in PRD.

Parties who would really want to interchange facts in RIF (instead of, 
e.g. XML data documents or whatever other means that has been designed 
for that purpose) can use BLD for that purpose.

> 3.4.1.
> Can you add to the editors note:
> A precise statement of the matching theory with respect to the data 
> types and builtins in DTB will be added.

I will.

> Throughout -- I would prefer American over British spellings, e.g. 
> labeled not labelled, serialized not serialised, etc.

I will check that a consistent spelling is used. I believe, indeed, that 
American spelling is the norm for W3C specs: I will try to check as far 
as I can.

> Fix as many as you can (and agree with) and then publish!


> I'll "see" you all at the July 22 telecon.

Have a nice vacation!


Received on Friday, 4 July 2008 08:20:02 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:47:51 UTC